EXT_SUFFIX too short on FreeBSD and AIX

[mattip]

BPO	42604
Nosy	@vstinner, @mattip, @pablogsal, @miss-islington
PRs	python/cpython#23708 python/cpython#23866 python/cpython#23867

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields: ```python assignee = None closed_at = created_at = labels = ['3.8', 'build', '3.7', '3.9', '3.10'] title = 'EXT_SUFFIX too short on FreeBSD and AIX' updated_at = user = 'https://github.com/mattip' ``` bugs.python.org fields: ```python activity = actor = 'vstinner' assignee = 'none' closed = True closed_date = closer = 'vstinner' components = ['Build'] creation = creator = 'mattip' dependencies = [] files = [] hgrepos = [] issue_num = 42604 keywords = ['patch'] message_count = 13.0 messages = ['382767', '383299', '383306', '383387', '383413', '383414', '383415', '383417', '385347', '385351', '385356', '385357', '385959'] nosy_count = 4.0 nosy_names = ['vstinner', 'mattip', 'pablogsal', 'miss-islington'] pr_nums = ['23708', '23866', '23867'] priority = 'normal' resolution = 'fixed' stage = 'resolved' status = 'closed' superseder = None type = None url = 'https://bugs.python.org/issue42604' versions = ['Python 3.7', 'Python 3.8', 'Python 3.9', 'Python 3.10'] ```

[mattip]

Continuation of bpo-39825, this time for FreeBSD and AIX. As commented there, the test added in the fix to 39825 fails on FreeBSD and AIX:

FAIL: test_EXT_SUFFIX_in_vars (test.test_sysconfig.TestSysConfig) ----------------------------------------------------------------------

Traceback (most recent call last):
  File "/usr/home/buildbot/python/3.x.koobs-freebsd-564d/build/Lib/test/test_sysconfig.py", line 368, in test_EXT_SUFFIX_in_vars
    self.assertEqual(vars['EXT_SUFFIX'], _imp.extension_suffixes()[0])
AssertionError: '.so' != '.cpython-310d.so'
- .so
+ .cpython-310d.so

So somehow EXT_SUFFIX is being set to .so rather than .cpython-310d.so.

It seems the difference in EXT_SUFFIX comes from this stanza in configure:

case $ac_sys_system in Linux|GNU|Darwin|VxWorks) EXT_SUFFIX=.${SOABI}${SHLIB_SUFFIX};; *) EXT_SUFFIX=${SHLIB_SUFFIX};; esac

where $ac_sys_system is uname -s. On FREEBSD, this is "FreeBSD", and I think on AIX it is "AIX". My preference would be to always set EXT_SUFFIX to ${SOABI}${SHLIB_SUFFIX}, with no option for setting it to a different value. Does that seem right?

[mattip]

vstinner: ping. Are merge builds still failing or can I close this?

[vstinner]

test_sysconfig still fails:

• POWER6 AIX 3.x: https://buildbot.python.org/all/#/builders/302/builds/563
• PPC64 AIX 3.x: https://buildbot.python.org/all/#/builders/438/builds/593
• AMD64 FreeBSD Shared 3.x: https://buildbot.python.org/all/#/builders/483/builds/572

[mattip]

The linked PR 23708 is meant to fix the failures. Any chance you can take a look?

[pablogsal]

New changeset a44ce6c9f725d336aea51a946b42769f29fed613 by Matti Picus in branch 'master': bpo-42604: always set EXT_SUFFIX=${SOABI}${SHLIB_SUFFIX} when using configure (GH-23708) https://github.com/python/cpython/commit/a44ce6c9f725d336aea51a946b42769f29fed613

[miss-islington]

New changeset b01091a3e71e6636d2df4db45920e820cdf7df3b by Miss Islington (bot) in branch '3.8': bpo-42604: always set EXT_SUFFIX=${SOABI}${SHLIB_SUFFIX} when using configure (GH-23708) https://github.com/python/cpython/commit/b01091a3e71e6636d2df4db45920e820cdf7df3b

[pablogsal]

New changeset 4b155967b3e743cbdc31600f13f1bfcf07f7b6ce by Miss Islington (bot) in branch '3.9': bpo-42604: always set EXT_SUFFIX=${SOABI}${SHLIB_SUFFIX} when using configure (GH-23708) (GH-23866) https://github.com/python/cpython/commit/4b155967b3e743cbdc31600f13f1bfcf07f7b6ce

[pablogsal]

Thanks for the PR, mattip!

[vstinner]

I reopen the issue.

It seems like Python no long builds on PPC64 AIX 3.x buildbot: https://buildbot.python.org/all/#/builders/438/builds/753

"setup.py build" crash with an assertion error:

Assertion failed: (item != NULL) ^ (PyErr_Occurred() != NULL), file Objects/abstract.c, line 163

The linker also displays tons of warnings. Examples:

ld: 0711-327 WARNING: Entry point not found: PyInituuid.cpython-310d ld: 0711-327 WARNING: Entry point not found: PyInitmultiprocessing.cpython-310d ld: 0711-327 WARNING: Entry point not found: PyInit__tkinter.cpython-310d

Can it be related to the commit a44ce6c9f725d336aea51a946b42769f29fed613 of this issue?

[pablogsal]

Can it be related to the commit a44ce6c9f725d336aea51a946b42769f29fed613 of this issue?

Quite likely. @mattip, can you investigate?

[vstinner]

Assertion failed: (item != NULL) ^ (PyErr_Occurred() != NULL), file Objects/abstract.c, line 163

This is bpo-42979: "_zoneinfo: zoneinfomodule_exec() doesn't check for PyDateTime_IMPORT failure".

It's a side effect of the broken _datetime module.

[mattip]

I think this issue can only cause failed tests. sysconfig.get_config_var('EXT_SUFFIX') is, as far as I can tell, not used internally by any cpython build or import code.

[vstinner]

It seems like Python no long builds on PPC64 AIX 3.x buildbot: https://buildbot.python.org/all/#/builders/438/builds/753

Sorry, it was unrelated to this issue. I fixed it with:

New changeset 0837f99d3367ecf200033bbddfa05d061ae9f483 by Victor Stinner in branch 'master': bpo-42323: Fix math.nextafter() on AIX (GH-24381)

This is bpo-42979: "_zoneinfo: zoneinfomodule_exec() doesn't check for PyDateTime_IMPORT failure".

_zoneinfo has also been fixed.